perplexity

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In the skill’s required runtime, the outsider-authored input/query (user text) is sent to Perplexity to perform web search and generate answers, so attacker-controlled text can influence retrieved outsider web content that the model consumes (via web_search tools and returned grounding snippets/answer text).